Overview
Cohen's d measures how big a difference between two group means is, in standard deviations. Unlike a p-value, it doesn't depend on the sample size.
Enter both group means, the pooled standard deviation and the total sample size. You also get Hedges' g, a version corrected for small samples.
Worked examples
| Inputs | Result |
|---|---|
| Group Mean 178Group Mean 272Pooled Std Dev10Total Sample Size84 | Cohen's d: 0.6 (medium)Hedges' g: 0.59Mean difference: 6 |
Show 2 more examples
| Group Mean 1105Group Mean 2100Pooled Std Dev15Total Sample Size60 | Cohen's d: 0.33 (small)Hedges' g: 0.33Mean difference: 5 |
| Group Mean 13.4Group Mean 22.1Pooled Std Dev1.2Total Sample Size30 | Cohen's d: 1.08 (large)Hedges' g: 1.05Mean difference: 1.3 |
Cohen's d by group mean 1
With the calculator's starting values, the cohen's d is 0.6. The table shows how it changes with the group mean 1.
| Group Mean 1 | Cohen's d |
|---|---|
| 40 | -3.2 |
| 50 | -2.2 |
| 60 | -1.2 |
| 70 | -0.2 |
| 78starting value | 0.6 |
| 80 | 0.8 |
| 90 | 1.8 |
| 100 | 2.8 |
| 110 | 3.8 |
Other inputs: Group Mean 2: 72 · Pooled Std Dev: 10 · Total Sample Size: 84. Change them in the calculator above for your own numbers.
How it works
d is the difference in means divided by the pooled standard deviation. As a rough guide, 0.2 is small, 0.5 medium and 0.8 large. Hedges' g multiplies d by 1 − 3 ÷ (4N − 9), which removes a slight bias in small samples.
The formula
| pooled SD | Combined standard deviation of both groups |
Practical tips
- Means of 78 and 72 with a pooled SD of 10 give d = 0.6, a medium effect.
- The pooled SD for equal groups is √((s₁² + s₂²) ÷ 2).
- Context matters. A 'small' effect can be important in medicine or education.
Frequently asked questions
What's the difference between d and g?
g corrects d for small-sample bias. They're nearly identical for large samples.
Can d be negative?
Yes, if group 2's mean is higher. The size is what matters.
How do I get the pooled SD?
√(((n₁ − 1)s₁² + (n₂ − 1)s₂²) ÷ (n₁ + n₂ − 2)).